Multiple Imputation for Recurent Event Endpoints in Clinical Trials
The dejaVu package performs multiple imputation on recurrent event data sets, following the approach described by Keene et al. The package can be used to perform multiple imputation of an existing study dataset where some patients dropped out. Imputation can be performed either assuming dropout is at random (missing at random) or assuming a specific non-random dropout mechanism (missing not at random). The package can also be used to simulate recurrent event datasets, in order to evaluate the impact of dropout and the properties of multiple imputation based analyses. Finally, the imputed data sets can be analysed and their results combined using Rubin’s rules.
Bartlett, Jonathan (maintainer); Burkoff, Nikolas; Metcalfe, Paul; Ruau, David;
To install the development version from GitHub:
install.packages("devtools") # We spent a lot of time developing the vignettes. We recommend the read but # building them from source takes some time devtools::install_github("scientific-computing-solutions/dejaVu", build_vignettes = TRUE)